NewYour coding agent can read the release notes before it upgrades.Set up the MCP server →
crates.io, #988 most downloaded on crates.io
A crate to help you go wide.
Latest stable
1.7.1
Last release 2 months ago
14 Sep 2026
Release timing varies
gaps range from 8 days to 3 months
Some releases are documented
notes for 16 of the last 60 stable releases
11 versions withdrawn
withdrawn after publishing
7 years old
87 releases · first in 2019
Drag across the ruler to compare, or tap a release to open it.
chore: Release wide version 1.7.1
chore: Release wide version 1.7.1
chore: Release wide version 1.7.0
chore: Release wide version 1.7.0
Select trait for arbitrary types in the select methodselect with mixed types (e.g., selecting f32x4 values
with u32x4 mask)from_i16x16 and From<i16x16> to i32x16.from_i8x32 and From<i8x32> to i16x32.swizzle and
swizzle_relaxed have been deprecated. See
PR #317 for more information.u8xL shifts by a constant on AVX by using shift instructions on
u16x(L/2), followed by masking out the bits shifted across bytes.
See PR #319, which supersedes
#313.from_ne_bytes and to_ne_bytes to all SIMD vector types. See
this issue for more
information.u8x64 and i8x64 types. See
this PR.signum documentation: the docs previously misstated the grammar
and the set of inputs that map to 1.0/-1.0. The behavior itself was
already correct and matches [f32::signum].signum. This changes the bit-patterns of returned NaNs.copysign.add_mul_lo and add_mul_hi for u32xN and u64xN: a fused
multiply-add over the low W bits of each lane, keeping the low or high half
of the product. u64xN uses vpmadd52lo/hi on AVX-512-IFMA when W == 52.unpack_lo and unpack_hi for u32xN and u64xN.Your coding agent can read these notes before it upgrades. Set up the MCP server →
chore: Release wide version 1.6.1
chore: Release wide version 1.6.1
avx512f was enabled without
avx512dq, which is a valid potential build. CI checks have been added for
this combination to prevent recurrence.chore: Release wide version 1.6.0
chore: Release wide version 1.6.0
i64xN functions to_bitmask, any, all and none
no longer guarantee that the sign-bit is used to decide whether an element is
considered "true" or "false". This is being counted as a "bug" in the docs for
previously being too specific.cast_unsigned, cast_signed, to_bits, from_bits.u8x32/i8x32 swizzle and swizzle_relaxed: a full-width 32-entry
byte table lookup (vpermb on AVX-512-VBMI, vqtbl2 on NEON, emulated on
AVX2/SSSE3).i8x32/u8x32 swizzle_half on AVX2: out-of-range indices now
correctly zero their output lane (previously leaked self[..][0]).to_array, as_array and as_mut_array available in const contexts.pow_{simd-type-name} to powf_simd and deprecated powf.wide types and native intrinsics SIMD types.reduce_mul for integers.reduce_mul and mul_keep_low_high.round_ties_even.any, all, none and fast_clamp.isize and usize and
SIMD vectors of "opposite-signedness".blend and replaced it with select and bitselect.powf.clamp and added guarantees for fast_clamp.saturating_add/sub.round which previously behaved like round_ties_even.UpperExp formatting for floats.mul_widen to widening_mul and added it for
remaining types.mul_keep_high for remaining types.#[must_use] annotationschore: Release wide version 1.5.0
chore: Release wide version 1.5.0
simd_* comparison functions and deprecated the Cmp* traits.Div and Rem implementations.clamp, saturating_mul and saturating_div.is_positive and signum.f64xN functions round and round_int.simd_ne NaN behavior for floats.u64x8::simd_lt.i64x8 and u64x8 function to_bitmasku32x16 functions any, all and none.is_negative.chore: Release wide version 1.4.0
chore: Release wide version 1.4.0
Rem for float types.sign_bit to is_sign_negative and added
is_sign_positive.chore: Release wide version 1.3.0
chore: Release wide version 1.3.0
f32x16 functions is_finite and round_int. They
previously gave incorrect output in some cases.f32x16 functionality.chore: Release wide version 1.2.0
chore: Release wide version 1.2.0
chore: Release wide version 1.1.1
chore: Release wide version 1.1.1
chore: Release wide version 1.1.0
chore: Release wide version 1.1.0
i64x2::min, i64x2::max.u64x2::min, u64x2::max.u64x8::min.chore: Release wide version 1.0.3
chore: Release wide version 1.0.3
chore: Release wide version 1.0.2
chore: Release wide version 1.0.2
wide was using the wrong avx512 sub-features to select
when safe_arch functions could be called, causing build errors.* initial stable version
Nothing published for this version
to_bitmask implemented for all unsigned int types.
to_bitmask implemented for all unsigned int types.Fixed additional build errors on wasm targets.
Fixed type errors in the fallback implementations of some methods that completely prevented compilation of the crate on some targets.
move_mask returns unsigned values now.
move_mask returns unsigned values now.move_mask and many other methods renamed to better align with the standard
library portable simd types. See issue
209 for details.Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
One column per quarter.